MapReduce particle filtering with exact resampling and deterministic runtime



Thiyagalingam, J ORCID: 0000-0002-2167-1343, Kekempanos, L and Maskell, S ORCID: 0000-0003-1917-2913
(2017) MapReduce particle filtering with exact resampling and deterministic runtime. Eurasip Journal on Advances in Signal Processing, 2017 (1).

This is the latest version of this item.

[img] Text
1705.01660v1.pdf - Submitted version

Download (790kB)
[img] Text
paper_all.pdf - Author Accepted Manuscript

Download (794kB)

Abstract

Particle filtering is a numerical Bayesian technique that has great potential for solving sequential estimation problems involving non-linear and non-Gaussian models. Since the estimation accuracy achieved by particle filters improves as the number of particles increases, it is natural to consider as many particles as possible. MapReduce is a generic programming model that makes it possible to scale a wide variety of algorithms to Big data. However, despite the application of particle filters across many domains, little attention has been devoted to implementing particle filters using MapReduce. In this paper, we describe an implementation of a particle filter using MapReduce. We focus on a component that what would otherwise be a bottleneck to parallel execution, the resampling component. We devise a new implementation of this component, which requires no approximations, has O(N) spatial complexity and deterministic O((logN)2) time complexity. Results demonstrate the utility of this new component and culminate in consideration of a particle filter with 224 particles being distributed across 512 processor cores.

Item Type: Article
Additional Information: 31 pages, 16 figures
Uncontrolled Keywords: MCMC methods, Particle filters, Big data sampling, MapReduce, Resampling
Depositing User: Symplectic Admin
Date Deposited: 11 Oct 2017 13:29
Last Modified: 19 Jan 2023 06:53
DOI: 10.1186/s13634-017-0505-9
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3009952

Available Versions of this Item